Advanced intelligent prediction method, system and equipment for dust concentration in limited space and medium

Abstract

The invention discloses an advanced intelligent prediction method, system and equipment for dust concentration in a limited space and a medium, and relates to the technical field of dust prevention and control. The method comprises the following steps: acquiring environmental parameters and current dust concentration in the confined space; preprocessing the environmental parameters to obtain preprocessed data; inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction to obtain a dust concentration prediction value with a time sequence; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism; and displaying the dust concentration predicted value in a graphical interface, and triggering an alarm when the dust concentration predicted value reaches a preset threshold value. According to the invention, the accuracy and advancement of dust concentration prediction in the limited space can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of dust prevention and control, and in particular to a method, system, equipment and medium for intelligently predicting dust concentration in a confined space. Background Art

[0002] In confined space working environments such as mines, accurate prediction of dust concentration is crucial to protecting worker health and production safety. Excessive dust concentration may lead to occupational diseases and environmental problems. Therefore, developing effective prediction methods is of great significance for preventing and controlling dust pollution.

[0003] Traditional dust concentration prediction methods, such as those based on physical models, are often limited in processing large data sets and have difficulty capturing the complexity and nonlinear characteristics of dust concentration changes. In recent years, deep learning techniques, especially recurrent neural networks (RNNs) and convolutional neural networks (CNNs), have been widely used in the fields of time series prediction and image recognition. These models are used to develop more accurate prediction models for pollutant concentrations such as PM2.5 due to their strong memory and feature extraction capabilities. Deep recurrent neural networks (DRNNs) combine the time series analysis capabilities of recurrent neural networks with the high-dimensional feature extraction capabilities of deep learning, and can better capture the dynamic characteristics and potential patterns of dust concentration changes. However, dust concentration in confined spaces is affected by many factors, including meteorological conditions, work activities, and spatial structures, and the results obtained using currently available traditional prediction methods are not accurate. Therefore, it is a technical challenge to develop an intelligent method that can integrate the influence of these factors and make advance predictions. Summary of the invention

[0004] The purpose of the present invention is to provide a method, system, device and medium for advanced intelligent prediction of dust concentration in a confined space, which can improve the accuracy and advanceness of dust concentration prediction in a confined space.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] An advanced intelligent prediction method for dust concentration in a confined space, comprising:

[0007] Obtain environmental parameters and current dust concentration in confined spaces;

[0008] Preprocessing the environmental parameters to obtain preprocessing data;

[0009] The pre-processed data and the current dust concentration are input into a trained deep recurrent neural network prediction module for prediction, so as to obtain a dust concentration prediction value with a time series; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism;

[0010] The dust concentration prediction value is displayed in a graphical interface, and an alarm is triggered when the dust concentration prediction value reaches a preset threshold.

[0011] Optionally, the preprocessing method includes cleaning, normalization and missing value filling to ensure the integrity and consistency of data input.

[0012] Optionally, before making a prediction, it also includes: extracting the environmental parameters to obtain data features; the data features include dust concentration historical data features, meteorological condition features, work activity data features, ventilation condition features, particle size and property features, geographical location and spatial structure features, image features, multimodal data fusion features, time series features, statistical features and frequency domain features.

[0013] Optionally, the training process of the deep recurrent neural network prediction module is:

[0014] Acquiring training data; the training data includes historical parameters and corresponding concentration prediction results;

[0015] Building a pre-trained network

[0016] The training data is input into the pre-trained network, and according to the set optimization algorithm, parameter optimization is performed with the goal of minimizing the loss between the network output and the concentration prediction result, and the network after parameter optimization is determined as the trained deep recurrent neural network prediction module.

[0017] Optionally, the setting optimization algorithm includes a gradient descent optimization algorithm, Bayesian optimization, regularization and gradient clipping; the gradient descent optimization algorithm adopts an Adam optimizer or a RMSProp optimizer.

[0018] The present invention also provides an advanced intelligent prediction system for dust concentration in a confined space, comprising:

[0019] Data acquisition unit, used to obtain environmental parameters and current dust concentration in the confined space;

[0020] A data preprocessing unit, used for preprocessing the environmental parameters to obtain preprocessed data;

[0021] A prediction unit, used for inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction, so as to obtain a dust concentration prediction value with a time series; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism;

[0022] The early warning unit is used to display the dust concentration prediction value in a graphical interface and trigger an alarm when the dust concentration prediction value reaches a preset threshold.

[0023] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned advanced intelligent prediction method for dust concentration in a confined space.

[0024] The present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned method for advanced intelligent prediction of dust concentration in a confined space.

[0025] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0026] The present invention discloses a method, system, device and medium for intelligently predicting dust concentration in a confined space. The method comprises obtaining environmental parameters and current dust concentration in the confined space; preprocessing the environmental parameters to obtain preprocessed data; inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction to obtain a dust concentration prediction value with a time series; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism; the dust concentration prediction value is displayed in a graphical interface, and an alarm is triggered when the dust concentration prediction value reaches a preset threshold. The present invention can improve the accuracy and advance of dust concentration prediction in confined spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0028] Figure 1 It is a schematic diagram of the process of the advanced intelligent prediction method of dust concentration in confined space of the present invention;

[0029] Figure 2 is a DRNN flow chart in this embodiment; wherein, part (a) is a first flow connection diagram; part (b) is a second flow connection diagram; and part (c) is a third flow connection diagram;

[0030] Figure 3 This is the dust concentration intelligent prediction diagram in this embodiment;

[0031] Figure 4 Schematic diagram of the LSTM unit structure in this embodiment;

[0032] Figure 5 Schematic diagram of the attention mechanism structure in this embodiment. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] The purpose of the present invention is to provide a method, system, device and medium for advanced intelligent prediction of dust concentration in a confined space, which can improve the accuracy and advanceness of dust concentration prediction in a confined space.

[0035] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] like Figure 1-Figure 5 As shown, the present invention provides an advanced intelligent prediction method for dust concentration in a confined space, comprising:

[0037] Step 100: Obtain environmental parameters and current dust concentration in the confined space.

[0038] Step 200: preprocess the environmental parameters to obtain preprocessed data, wherein the preprocessing method includes cleaning, normalization and missing value filling to ensure the integrity and consistency of data input.

[0039] Step 300: Input the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction to obtain a dust concentration prediction value with a time series; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism.

[0040] Step 400: Display the dust concentration prediction value in a graphical interface, and trigger an alarm when the dust concentration prediction value reaches a preset threshold.

[0041] As a specific implementation method, before making a prediction, it also includes: extracting the environmental parameters to obtain data features; the data features include dust concentration historical data features, meteorological condition features, work activity data features, ventilation condition features, particle size and property features, geographical location and spatial structure features, image features, multimodal data fusion features, time series features, statistical features and frequency domain features.

[0042] As a specific implementation, the training process of the deep recurrent neural network prediction module is:

[0043] Acquiring training data; the training data includes historical parameters and corresponding concentration prediction results;

[0044] Building a pre-trained network

[0045] The training data is input into the pre-trained network, and according to the set optimization algorithm, the parameters are optimized with the goal of minimizing the loss between the network output and the concentration prediction result, and the network after parameter optimization is determined as the trained deep recurrent neural network prediction module. The set optimization algorithm includes a gradient descent optimization algorithm, Bayesian optimization, and regularization and gradient clipping; the gradient descent optimization algorithm uses an Adam optimizer or an RMSProp optimizer.

[0046] Based on the above technical solution, the following embodiments are provided.

[0047] System Overview:

[0048] This technical solution aims to develop an intelligent prediction system based on deep recurrent neural network (DRNN) for dust concentration prediction in confined spaces (such as mines, tunnels, warehouses, etc.). Through real-time data collection, preprocessing and intelligent prediction modules, the system provides advanced warning for environmental safety monitoring and dust concentration control, reduces accident risks, and improves health and safety in the workplace.

[0049] System composition:

[0050] (1) Data acquisition module: Equipped with a variety of sensors (such as dust concentration, temperature and humidity, air flow rate sensors, etc.) to monitor environmental parameters in confined spaces in real time. The collected data is transmitted to the data processing center via wireless transmission.

[0051] Feature extraction: Feature extraction is a crucial step when building a dust concentration prediction model based on deep recurrent neural networks. The purpose of feature extraction is to convert raw data into a format that the model can understand and process, while capturing key information about the changes in dust concentration. Here are some key features to extract: 1) Dust concentration history data: This is the most direct feature, including dust concentration measurements at different time points in the past. This data can help the model learn the changing trend of dust concentration over time. 2) Meteorological conditions: Including temperature, humidity, wind speed and wind direction, these factors may affect the diffusion and concentration of dust. 3) Operation activity data: Operation activities in the mine, such as blasting, excavation and transportation, can also affect the generation and distribution of dust. 4) Ventilation conditions: The ventilation system in the mine, including wind speed, air volume and the configuration of the ventilation network, has a significant impact on the diffusion of dust. 5) Particle size and properties: The particle size distribution and chemical properties in the dust also affect its behavior in the air. 6) Geographic location and spatial structure: Different areas of the mine may have different dust concentration distributions, so geographic location and spatial structure are also important features. 7) Image features: If image analysis methods are used to detect dust, relevant parameters such as grayscale features, texture features, and geometric features can be extracted from the image. 8) Multimodal data fusion: Different types of data (such as sensor data, image data, meteorological data, etc.) may need to be fused together to provide more comprehensive dust concentration features. 9) Time series features: Deep recurrent neural networks are particularly suitable for processing time series data, so features such as trends, seasonality, and periodicity can be extracted from time series. 10) Statistical features: Statistical analysis is performed on dust concentration data to extract statistical features such as mean, standard deviation, maximum, and minimum. 11) Frequency domain features: Use methods such as Fourier transform to convert time series data to the frequency domain and extract frequency domain features. Through these feature extraction methods, rich information can be provided to the deep recurrent neural network model to help the model more accurately predict changes in dust concentration.

[0052] (2) Data preprocessing module: cleans, normalizes and fills missing values ​​in the collected data, replaces the data (interpolation method), and ensures the integrity and consistency of the data input.

[0053] (3) Deep recurrent neural network prediction module: A deep recurrent neural network (DRNN) model with an attention mechanism-based LSTM (long short-term memory network) or GRU (gated recurrent unit) structure is used to predict the time series of dust concentration. The model input includes the current dust concentration and environmental parameters, and the model output is the predicted value of dust concentration in the future. When constructing a dust concentration prediction model based on a deep recurrent neural network, the following optimization algorithms are used to improve the model's convergence speed and robustness, generalization ability, and prediction performance:

[0054] 1) Gradient descent optimization algorithm

[0055] Adam optimizer: combines the advantages of momentum and adaptive learning rate adjustment, suitable for training deep learning models, especially tasks with non-stationary time series.

[0056] RMSProp optimizer: used to process noise in sequence data, can stably update the learning rate, and is suitable for time series tasks in DRNN.

[0057] 2) Bayesian Optimization

[0058] In hyperparameter tuning, Bayesian optimization can be used to select the best combination of model parameters, including the number of hidden units, learning rate, regularization parameters, etc.

[0059] 3) Regularization and gradient clipping

[0060] In constrained spatial data, there may be gradient explosion or vanishing problems. Using methods such as gradient clipping and weight decay can enhance the stability of the model.

[0061] (4) Prediction result display and alarm module: The prediction results are displayed through a graphical interface, and an alarm is triggered when the dust concentration reaches the preset threshold so that preventive measures can be taken.

[0062] Therefore, this embodiment develops an advanced intelligent prediction system based on deep recurrent neural network (DRNN) for predicting dust concentration in confined spaces (such as mines, tunnels, warehouses, etc.). This technology is suitable for confined spaces such as mines, tunnel construction, chemical workshops, etc. that require dust concentration monitoring, and can be applied to labor protection, environmental monitoring and safety management. It has the following beneficial effects:

[0063] (1) Intelligent prediction: DRNN’s time series modeling capability can accurately predict the changing trend of dust concentration, providing a basis for taking safety measures in advance.

[0064] (2) Real-time monitoring and advance warning: Data is collected and processed in real time, and the prediction results can be responded to in seconds, which is suitable for confined spaces where dust concentration fluctuates frequently.

[0065] (3) Strong adaptability: The system can make adaptive adjustments based on historical data of different environmental conditions and is suitable for a variety of confined space application scenarios.

[0066] In addition, the present invention also provides an advanced intelligent prediction system for dust concentration in a confined space, comprising:

[0067] Data acquisition unit, used to obtain environmental parameters and current dust concentration in the confined space;

[0068] A data preprocessing unit, used for preprocessing the environmental parameters to obtain preprocessed data;

[0069] A prediction unit, used for inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction, so as to obtain a dust concentration prediction value with a time series; the deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism;

[0070] The early warning unit is used to display the dust concentration prediction value in a graphical interface and trigger an alarm when the dust concentration prediction value reaches a preset threshold.

[0071] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned advanced intelligent prediction method for dust concentration in a confined space.

[0072] The present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned method for advanced intelligent prediction of dust concentration in a confined space.

[0073] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0074] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for advanced intelligent prediction of dust concentration in confined spaces, characterized in that: include: Obtain environmental parameters and current dust concentration in confined spaces; Preprocessing the environmental parameters to obtain preprocessing data; Inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction to obtain a dust concentration prediction value with a time series; The deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism; The dust concentration prediction value is displayed in a graphical interface, and an alarm is triggered when the dust concentration prediction value reaches a preset threshold.

2. The method for advanced intelligent prediction of dust concentration in confined spaces according to claim 1 is characterized in that: The preprocessing methods include cleaning, normalization and missing value filling to ensure the integrity and consistency of data input.

3. The method for advanced intelligent prediction of dust concentration in confined spaces according to claim 1 is characterized in that: Before making a prediction, it also includes: extracting the environmental parameters to obtain data features; the data features include dust concentration historical data features, meteorological condition features, work activity data features, ventilation condition features, particle size and property features, geographical location and spatial structure features, image features, multimodal data fusion features, time series features, statistical features and frequency domain features.

4. The method for advanced intelligent prediction of dust concentration in confined spaces according to claim 1 is characterized in that: The training process of the deep recurrent neural network prediction module is: Acquiring training data; the training data includes historical parameters and corresponding concentration prediction results; Building a pre-trained network The training data is input into the pre-trained network, and according to the set optimization algorithm, parameter optimization is performed with the goal of minimizing the loss between the network output and the concentration prediction result, and the network after parameter optimization is determined as the trained deep recurrent neural network prediction module.

5. The method for advanced intelligent prediction of dust concentration in confined spaces according to claim 4 is characterized in that: The setting optimization algorithm includes gradient descent optimization algorithm, Bayesian optimization and regularization and gradient clipping; the gradient descent optimization algorithm adopts Adam optimizer or RMSProp optimizer.

6. An advanced intelligent prediction system for dust concentration in confined spaces, characterized in that: include: Data acquisition unit, used to obtain environmental parameters and current dust concentration in the confined space; A data preprocessing unit, used for preprocessing the environmental parameters to obtain preprocessed data; A prediction unit, used for inputting the preprocessed data and the current dust concentration into a trained deep recurrent neural network prediction module for prediction, so as to obtain a dust concentration prediction value with a time series; The deep recurrent neural network prediction module adopts a deep recurrent neural network model based on an attention mechanism; The early warning unit is used to display the dust concentration prediction value in a graphical interface and trigger an alarm when the dust concentration prediction value reaches a preset threshold.

7. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the method for advanced intelligent prediction of dust concentration in a confined space according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements the method for advanced intelligent prediction of dust concentration in a confined space as described in any one of claims 1 to 5.

Images